Blaizzy

diff --git a/src/transformers/models/csm/convert_csm.py b/src/transformers/models/csm/convert_csm.py
index 28fbc9fe49..fab38f5ec5 100644
--- a/src/transformers/models/csm/convert_csm.py
+++ b/src/transformers/models/csm/convert_csm.py
@@ -31,6 +31,7 @@ from transformers import (
     MimiModel,
 )
 from transformers.utils.hub import cached_file
+from safetensors.torch import load_file, save_file
 

Blaizzy

import argparse
import asyncio
import json
import logging
import os
import wave

import aiohttp
import numpy as np
import sounddevice as sd

Blaizzy

import json
import time
from dataclasses import dataclass
from pathlib import Path
from typing import Any, Dict, Optional, Union

import mlx.core as mx
import mlx.nn as nn
from mlx_lm.generate import stream_generate
from mlx_lm.models.base import BaseModelArgs, create_attention_mask

Blaizzy

import argparse
import gradio as gr
import requests
import json

import asyncio


async def process_sse_stream(url, headers, data):
    response = requests.post(url, headers=headers, json=data, stream=True)

Blaizzy

import aiohttp
import json
import gradio as gr
from gradio import ChatMessage



class CustomChatAPI:
    def __init__(self, base_url, model):
        self.base_url = base_url

Blaizzy

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from datasets import load_dataset
from torch.utils.data import DataLoader
import os
import pyarrow as pa
import pyarrow.parquet as pq
from tqdm import tqdm

def process_and_save_parquet(all_logits, all_input_ids, all_attention_masks, top_50_token_ids, top_50_values, output_dir, shard_idx, total_shards):

Blaizzy

import torch
import torch.nn.functional as F
from transformers import AutoModelForCausalLM, AutoTokenizer, Trainer, TrainingArguments
from trl import SFTTrainer, SFTConfig
from datasets import load_dataset
import numpy as np

# Hyperparameters
temperature = 0.7
alpha = 0.5 

Blaizzy

base_model: meta-llama/Meta-Llama-3-70B
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: /home/migel/ai_datasets/tess-v1.5b-chatml.jsonl

Blaizzy

base_model: mistral-community/Mixtral-8x22B-v0.1
model_type: MixtralForCausalLM
tokenizer_type: AutoTokenizer
is_mistral_derived_model: false
trust_remote_code: true

load_in_8bit: false
load_in_4bit: true
strict: false

Blaizzy

```python
from trl import SFTTrainer, DataCollatorForCompletionOnlyLM
from transformers import AutoTokenizer
from datasets import load_dataset

# Load Dataset and tokenizer
dataset = load_dataset('prince-canuma/tinyOrca', split='train')
tokenizer = AutoTokenizer.from_pretrained("prince-canuma/Damysus-2.7B-Chat")